Vehicle Control Apparatus

The present disclosure relates to a vehicle control apparatus that performs a vehicle control for changing a traveling state of a vehicle in order to reduce a collision risk when the collision risk is equal to or greater than a threshold risk.

Patent Document 1: Japanese Patent Application Laid-open No. 2008-024108 Conventionally, a vehicle control apparatus that performs a vehicle control when a collision risk with an object is equal to or greater than a threshold risk has been known. For example, a vehicle control apparatus described in Patent Document 1 (hereinafter referred to as “the conventional apparatus”) estimates, when a collision with the object is inevitable, collision damage representing damage to an occupant caused by the collision based on an amount of cabin deformation and a boarding state of the occupant in the vehicle. The conventional apparatus then performs the vehicle control based on a vehicle control amount for minimizing the collision damage.

When the vehicle collides with an object (hereinafter referred to as “a collision object”), damage occurs not only to the occupant of the vehicle but also to the collision object. Furthermore, the damage to the occupant may occur not only due to the collision damage to the occupant caused by the collision, but also due to a change in a traveling state of the vehicle caused by execution of the vehicle control. Since the conventional apparatus does not take into consideration the damage to the collision object and the damage to the occupant during the execution of the vehicle control, the conventional apparatus may not be able to perform the vehicle control appropriately.

The present disclosure is made to address the above problem. That is, one of the objects of the present invention is to provide a vehicle control apparatus capable of increasing a possibility of performing the vehicle control appropriately in consideration of the damage to the occupant caused by the collision, the damage to the occupant during the execution of the vehicle control, and the damage to the collision object caused by the collision.

420 450 A vehicle control apparatus according to the present disclosure (hereinafter referred to as “the present apparatus”), when a collision risk, which represents a risk of a vehicle colliding with an object, is equal to or greater than a threshold risk (step: “Yes”), performs vehicle control to change a traveling state of the vehicle according to a control pattern selected from among a plurality of control patterns so as to reduce the collision risk (step).

525 a first occupant damage (D1) representing damage to an occupant caused by a change in the traveling state during a vehicle control period from when the vehicle control is executed according to the control pattern until the vehicle collides with a collision object (step), 530 a second occupant damage (D2) representing damage to the occupant at a collision time when the vehicle collides with the collision object as a result of executing the vehicle control according to the control pattern (step), and 535 an object damage (D3) representing damage to the collision object at the collision time (step); and estimate, for each of the control patterns, 440 450 execute the vehicle control according to the control pattern that minimizes a total damage (Dt) acquired based on the first occupant damage, the second occupant damage, and the object damage (step, step). The present apparatus is configured to:

According to the present apparatus, the vehicle control is performed according to the control pattern that minimizes the total damage acquired based on the first occupant damage, the second occupant damage, and the object damage. Accordingly, the present apparatus can perform the vehicle control appropriately based on the first occupant damage, the second occupant damage, and the object damage.

10 10 20 20 20 1 FIG. A vehicle control apparatusaccording to an embodiment of the present disclosure (hereinafter, referred to as “the present apparatus”) is applied to a vehicle VA and comprises components shown in. In the present specification, an “ECU” is an electronic control unit having a microcomputer as a main component. The ECUis also referred to as a control unit, a controller and a computer. The microcomputer includes a CPU (processor), a ROM, a RAM and an interface (I/F), etc. Functions realized by the ECUmay be realized by multiple ECUs.

22 20 22 24 20 22 24 A cameraacquires forward image data by capturing an area in front of the vehicle VA. The ECUobtains the forward image data from the camera. A millimeter-wave radardetects an object present in front of the vehicle VA and acquires object data related to the object. The ECUobtains the forward image data and the object data from the front cameraand the millimeter-wave radar, respectively, and specifies a position of the object relative to the vehicle VA and a relative speed of the object with respect to the vehicle VA based on the forward image data and the object data.

26 28 20 26 28 30 20 30 A seatbelt wearing sensordetects a wearing state indicating whether an occupant of the vehicle VA is wearing a seatbelt (occupant restraint device). A seat sensordetects a sitting state indicating whether the occupant is seated on a seat of the vehicle VA. The ECUobtains detection values representing the states detected by these sensorsand. An in-vehicle cameraacquires in-vehicle image data by capturing a passenger compartment of the vehicle VA. The ECUobtains the in-vehicle image data from the in-vehicle cameraand specifies an occupant's posture, a gripping state indicating whether the occupant is gripping a structure in the passenger compartment (such as a handrail or a strap), the occupant's position in the passenger compartment, and the congestion state in the passenger compartment.

32 34 36 20 32 36 An acceleration sensormeasures a longitudinal acceleration Gx of the vehicle VA and a lateral acceleration Gy of the vehicle VA. The acceleration Gx in the forward direction of the vehicle VA is a positive value, and the acceleration Gx in the rearward direction of the vehicle VA is a negative value. The acceleration Gx having a negative value may be referred to as “a deceleration.” A vehicle speed sensormeasures a vehicle speed Vs representing a speed of the vehicle VA. A steered angle sensormeasures a steered angle θ of steered wheels. The ECUobtains measured values from the sensorsto.

38 38 a A communication interface (I/F)communicates with a servervia a network NW.

42 44 46 48 48 46 20 A powertrain actuatorchanges a driving force generated by a drive unit of the vehicle VA (for example, an internal combustion engine and/or an electric motor). A brake actuatorchanges a braking force applied to the vehicle VA. A steering motoris incorporated in a steering mechanism. The steering mechanismis a mechanism for steering the steered wheels according to a steering operation of a steering wheel (not shown) by the driver. Furthermore, the steering motorapplies an automatic steering torque for changing the steered angle θ (a traveling direction of the vehicle VA) of the steered wheels in accordance with an instruction from the ECU.

20 10 The ECUof the present apparatusperforms a vehicle control for changing a traveling state of the vehicle VA in order to reduce a collision risk representing a possibility that the vehicle VA collides with the object when the collision risk is equal to or greater than a threshold risk. Such vehicle control is a type of autonomous driving for assisting the driver's driving.

20 2 FIG. 3 FIG. Specifically, the ECUselects one control pattern each from a deceleration control pattern shown inand a steering control pattern shown in, and performs the vehicle control according to the selected control patterns. A plurality of deceleration control patterns with different decelerations set as target acceleration Gxtgt are prepared in advance. Similarly, a plurality of steering control patterns with different lateral accelerations set as target lateral acceleration Gytgt are prepared in advance.

20 20 The ECUestimates control occupant damage D1, collision occupant damage D2, and object damage D3 for all combinations of the deceleration control patterns and the steering control patterns, and acquires total damage Dt based on the control occupant damage D1, the collision occupant damage D2, and the object damage D3. The ECUperforms the vehicle control according to the combination of control patterns that minimizes the total damage Dt.

The control occupant damage D1 represents damage to the occupant caused by a change in the vehicle's traveling state while the vehicle control is being performed according to the above combination of the control patterns. The control occupant damage D1 may also be referred to as “first occupant damage.”

The collision occupant damage D2 represents damage to the occupant caused by the collision between the vehicle VA and a collision object while the vehicle control is being performed according to the above combination of the control patterns. The collision object represents an object with which the vehicle VA collides. The collision object may be different from the object whose collision risk was equal to or greater than the threshold risk at the start of the vehicle control, since the steering control performed as the vehicle control is intended to change the traveling direction of the vehicle VA. The collision occupant damage D2 may also be referred to as “second occupant damage.”

The object damage D3 represents damage to the collision object caused by the collision between the vehicle VA and the collision object.

As a result, the present apparatus 10 can increase the possibility of appropriately performing the vehicle control by considering not only the collision occupant damage D2 but also the control occupant damage D1 and the object damage D3.

2 FIG. 1 3 1 1 2 2 3 3 As shown in, three deceleration control patterns Bto Bare prepared in advance. In deceleration control pattern B, the target acceleration Gxtgt is set to “0.” That is, the deceleration control pattern Bis a control pattern that does not decelerate the vehicle VA. In deceleration control pattern B, the target acceleration Gxtgt is set to Gx1, which is a negative value. That is, the deceleration control pattern Bis a control pattern that gently decelerates the vehicle VA. In deceleration control pattern B, the target acceleration Gxtgt is set to Gx2, which is a negative value smaller than Gx1. That is, the deceleration control pattern Bis a control pattern that decelerates the vehicle VA rapidly.

3 FIG. 1 2 1 1 2 2 As shown in, two steering control patterns Sand Sare prepared in advance. In steering control pattern S, the target lateral acceleration Gytgt is set to “0.” That is, the steering control pattern Sis a control pattern that does not change the traveling direction of the vehicle VA. In steering control pattern S, the target lateral acceleration Gytgt is set to Gy1. That is, the steering control pattern Sis a control pattern that gently changes the traveling direction of the vehicle VA.

20 4 FIG. The CPU of the ECUexecutes a routine shown in a flowchart ofeach time a predetermined time has elapsed.

400 405 4 FIG. When an appropriate timing comes, the CPU starts a process from stepin. At step, the CPU determines whether or not an execution flag Xexe is “0”. The execution flag Xexe is set to “1” when the vehicle control is performed, and set to “0” when vehicle control is not performed. The execution flag Xexe is set to “0” in an initialization routine. The initialization routine is executed by the CPU when an ignition key switch (not shown) of the vehicle VA is switched from an off position to an on position.

405 410 420 If the execution flag Xexe is “0”, the CPU makes a “Yes” determination at stepand executes stepsto.

410 Step: The CPU obtains the forward image data and the object data.

415 Step: The CPU recognizes the object based on the forward image data and the object data, and acquires a TTC representing the time until the object collides with the vehicle VA. TTC stands for “Time To Collision.” TTC is an index value representing the collision risk of the object. The smaller the TTC, the higher the collision risk. The CPU acquires the TTC by dividing the distance between the vehicle VA and the object by the relative speed of the object with respect to the vehicle VA.

420 495 495 If the TTC is greater than a threshold time Tth, that is, if the collision risk is less than the threshold risk, the CPU makes a “No” determination at step, and the process proceeds to step. At step, the CPU terminates this routine tentatively.

420 425 435 425 Step: The CPU selects one combination of the deceleration control pattern and the steering control pattern. If the TTC is equal to or less than the threshold time Tth, that is, if the collision risk is equal to or greater than the threshold risk, the CPU makes a “Yes” determination at stepand executes stepsto.

430 425 Step: The CPU executes a total damage estimation subroutine for estimating the total damage Dt that is caused when the vehicle control is performed according to the combination of control patterns selected at step. Details of the total damage estimation subroutine will be described later.

435 Step: The CPU determines whether or not all combinations have been selected.

435 425 If all combinations have not been selected, the CPU makes a “No” determination at step. In this case, the process returns to step, and the CPU selects another combination of the deceleration control pattern and the steering control pattern that has not yet been selected.

435 440 445 On the other hand, if all combinations have been selected, the CPU makes a “Yes” determination at stepand executes stepsand.

440 Step: The CPU selects, as an execution pattern, the combination in which the total damage Dt is minimized.

445 Step: The CPU sets the execution flag Xexe to “1.”

495 Thereafter, the process proceeds to step.

405 405 450 455 If the execution flag Xexe is “1” when the process proceeds to step, the CPU makes a “No” determination at stepand executes stepsand.

450 42 44 46 42 44 46 Step: The CPU controls the powertrain actuator, the brake actuator, and the steering motorbased on the execution pattern. Specifically, the CPU controls the powertrain actuatorand the brake actuatorsuch that the longitudinal acceleration Gx of the vehicle VA matches the target longitudinal acceleration Gxtgt defined by the deceleration control pattern. The CPU also controls the steering motorsuch that the lateral acceleration Gy of the vehicle VA matches the target lateral acceleration Gytgt defined by the steering control pattern.

455 Step: The CPU determines whether or not an end condition for the vehicle control is satisfied. For example, the CPU determines that the end condition is satisfied when the collision risk of the object no longer exists or when the vehicle VA has stopped.

455 495 455 460 465 If the end condition is not satisfied, the CPU makes a “No” determination at step, and the process proceeds to step. If the end condition is satisfied, the CPU makes a “Yes” determination at step, and executes stepsand.

460 38 a Step: The CPU transmits control result information to the server. The acceleration generated during the vehicle control, the change in the occupant's posture during the vehicle control, the actual collision pattern, the change in the occupant's posture at the time of collision, and the damage to the actual collision object are specified based on the control result information.

It should be noted that, if the vehicle VA does not collide with the object, it is sufficient that the acceleration generated during the vehicle control and the change in the occupant's posture during the vehicle control can be specified based on the control result information.

For example, the CPU can specify the change in the occupant's posture during the vehicle control and at the time of collision based on the in-cabin image data acquired during the vehicle control and at the time of collision. The CPU can specify the collision pattern and the damage to the collision object based on the forward image data acquired at the time of the collision.

465 Step: The CPU sets the execution flag Xexe to “0.”

495 Thereafter, the process proceeds to step.

430 500 505 540 5 FIG. When the process proceeds to step, the CPU starts a process from stepinand executes stepsto.

505 Step: The CPU specifies an attribute of the occupant.

The attribute of the occupant includes the occupant's age and gender. The CPU specifies the occupant's age and gender based on information registered in the occupant's smartphone, transportation IC card, or the like. The CPU may alternatively specify the occupant's age and gender based on the in-vehicle image data.

510 Step: The CPU specifies a boarding state of the occupant.

26 28 The boarding state of the occupant includes the occupant's posture, the occupant's gripping state of the structure, the occupant's seatbelt wearing state, the occupant's position in the passenger compartment, and the congestion state in the passenger compartment. The CPU specifies the occupant's posture, the occupant's gripping state of the structure, the occupant's position in the passenger compartment, and the congestion state in the passenger compartment based on the in-vehicle image data. The CPU specifies the occupant's seatbelt wearing state based on the detection value of the seatbelt wearing sensor. The CPU may alternatively specify the occupant's position in the passenger compartment based on the detection value of the seat sensor.

515 Step: The CPU specifies an attribute of the collision object.

The attribute of the collision object is a type of the collision object. For example, the type of the collision object indicates whether the collision object is a vehicle, a pedestrian, a guardrail, a wall, or the like. The CPU specifies the type of the collision object based on the forward image data.

520 Step: The CPU estimates the collision pattern between the vehicle VA and the collision object.

The collision pattern includes a collision speed, which represents the relative speed of the collision object with respect to the vehicle VA at the time of the collision, and a collision position on the vehicle body of the vehicle VA where the collision object collides. The CPU specifies the collision speed based on the current vehicle speed Vs, the target accelerations (Gxtgt and Gytgt) defined by the selected control patterns, and the current relative speed of the collision object. Furthermore, the CPU estimates a future traveling route of the vehicle VA based on the current vehicle speed Vs and the target accelerations (Gxtgt and Gytgt) defined by the selected control patterns, and estimates a future traveling route of the collision object based on a “history of a position of the collision object with respect to the vehicle VA.” Then, the CPU specifies the collision position on the vehicle body based on the estimated future predicted traveling route of the vehicle VA and the estimated future predicted traveling route of the collision object.

525 Step: The CPU estimates the control occupant damage D1 based on the attribute of the occupant, the boarding state of the occupant, and a control acceleration.

The control acceleration represents an acceleration generated in the vehicle VA by the vehicle control executed according to the selected combination. The CPU specifies the control acceleration based on the current vehicle speed Vs and the target accelerations (Gxtgt and Gytgt) defined by the selected control patterns.

The control occupant damage D1 increases as the number of the occupants determined, based on their ages, to be infants or elderly persons increases. The control occupant damage D1 increases as the number of female occupants increases.

The control occupant damage D1 increases as the number of the occupants not seated increases. The control occupant damage D1 increases as the number of the occupants not gripping the structure increases. The control occupant damage D1increases as the number of the occupants not wearing the seatbelt increases. The control occupant damage D1 increases as the congestion in the passenger compartment increases.

530 Step: The CPU estimates the collision occupant damage D2 based on the attribute of the occupant, the boarding state of the occupant, the attribute of the collision object, and the collision pattern. The estimation of the collision occupant damage D2 based on the attribute of the occupant and the boarding state of the occupant is the same as that of the control occupant damage D1, and thus, a description thereof will be omitted. The closer the position of the occupant in the passenger compartment is to the collision position, the greater the collision occupant damage D2 becomes.

535 Step: The CPU estimates the object damage D3 based on the attribute of the collision object and the collision pattern.

540 Step: The CPU estimates the total damage Dt by applying the control occupant damage D1, the collision occupant damage D2, and the object damage D3 to the following equation (1).

Ga1, Ga2, and Ga3 are gains set to desired values in a range of “0.0” or more and “1.0” or less. In the present embodiment, Ga1, Ga2, and Ga3 are set to “1.0.”

595 435 4 FIG. Thereafter, the process proceeds to step, the CPU terminates the present routine tentatively, and the process proceeds to stepshown in.

10 Accordingly, the present apparatuscan increase the possibility of performing the vehicle control appropriately by taking into account not only the collision occupant damage D2 but also the control occupant damage D1 and the object damage D3.

6 FIG. 1 1 1 1 (1) When a combination of the deceleration control pattern Band the steering control pattern Sis selected, the vehicle VA travels straight along a route Rwithout decelerating. When the vehicle VA travels straight along the route R, the vehicle VA collides with a pedestrian PD. In the example shown in, the control occupant damage D1, the collision occupant damage D2, and the object damage D3 in each combination will be described.

2 1 (2) When a combination of the deceleration control pattern Band the steering control pattern Sis selected, since the vehicle VA decelerates gradually, the magnitude of acceleration generated by the vehicle control becomes greater than that in (1) above. Therefore, the control occupant damage D1 becomes greater than that in (1) above, namely “1 point.” Since the vehicle VA collides with the collision object after gradually decelerating, the collision vehicle speed becomes lower than that in (1) above. Furthermore, the type of the collision object is the same as in (1) above, namely “pedestrian.” Therefore, the collision occupant damage D2 becomes smaller than that in (1) above, namely “2 points,” and the object damage D3 becomes smaller than that in (1) above, namely “7 points.” Accordingly, the total damage Dt is 10 points. 3 1 (3) When a combination of the deceleration control pattern Band the steering control pattern Sis selected, since the vehicle VA decelerates rapidly, the magnitude of acceleration generated by the vehicle control becomes greater than that in (2) above. Therefore, the control occupant damage D1 is “2 points.” Since the vehicle VA collides with the collision object after rapidly decelerating, the collision vehicle speed becomes lower than that in (2) above. Furthermore, the type of the collision object is the same as in (1) and (2) above, namely “pedestrian.” Therefore, the collision occupant damage D2 becomes lower than that in (2) above, namely “1 point,” and the object damage D3 becomes lower than that in (1) above, namely “5 points.” Accordingly, the total damage Dt is 8 points. 1 2 2 2 (4) When a combination of the deceleration control pattern Band the steering control pattern Sis selected, the vehicle VA turns to the right along a route Rwithout decelerating. When the vehicle VA travels along the route R, the vehicle VA collides with a guardrail GR instead of the pedestrian PD. Since the vehicle VA turns to the right without decelerating, the magnitude of acceleration generated by the vehicle control becomes greater than that in (1) above. Therefore, the control occupant damage D1 is “1 point.” Since the vehicle VA collides with the collision object (guardrail GR) without decelerating, the collision vehicle speed is high. Furthermore, the type of the collision object is “guardrail.” Therefore, the collision occupant damage D2 is “6 points,” and the object damage D3 is “0 points.” Accordingly, the total damage Dt is 7 points. Since the vehicle VA travels straight without decelerating, no acceleration is generated by the vehicle control. Therefore, the control occupant damage D1 is “0 points.” Since the vehicle VA collides with the collision object (pedestrian PD) without decelerating, the collision vehicle speed is high. Furthermore, the type of the collision object (pedestrian PD) is “pedestrian.” Therefore, the collision occupant damage D2 is “3 points,” and the object damage D3 is “10 points.” Accordingly, the total damage Dt is 13 points.

2 2 (5) When a combination of the deceleration control pattern Band the steering control pattern Sis selected, since the vehicle VA turns to the right while gradually decelerating, the magnitude of acceleration generated by the vehicle control becomes greater than that in (2) above. Therefore, the control occupant damage D1 becomes greater than that in (2) above, namely “2 points.” Since the vehicle VA collides with the collision object after gradually decelerating, the collision vehicle speed becomes lower than that in (4) above. Furthermore, the type of the collision object is the same as in (4) above, namely “guardrail.” Therefore, the collision occupant damage D2 becomes smaller than that in (4) above, namely “4 points,” and the object damage D3 is “0 points.” Accordingly, the total damage Dt is 6 points. 3 2 (6) When a combination of the deceleration control pattern Band the steering control pattern Sis selected, since the vehicle VA turns to the right while decelerating rapidly, the magnitude of acceleration generated by the vehicle control becomes greater than that in (3) above. Therefore, the control occupant damage D1 is “3 points.” Since the vehicle VA collides with the collision object after rapidly decelerating, the collision vehicle speed becomes lower than that in (5) above. Furthermore, the type of the collision object is the same as in (4) and (5) above, namely “guardrail.” Therefore, the collision occupant damage D2 becomes lower than that in (2) above, namely “2 points,” and the object damage D3 becomes lower than that in (1) above, namely “0 points.” Accordingly, the total damage Dt is 5 points. The collision occupant damage D2 becomes lower when the type of the collision object is a pedestrian, and becomes higher when the type of the collision object is a vehicle, a guardrail, a wall, or the like. The object damage D3 becomes higher when the type of the collision object is the pedestrian, and becomes lower when the type of the collision object is the guardrail, the wall, or the like. In addition, when the type of the collision object is the vehicle, there is a possibility that a person is on board the vehicle, and thus, the object damage D3 becomes higher than when the type of the collision object is the guardrail, the wall, or the like.

3 2 3 2 Accordingly, since the total damage Dt becomes smallest when the deceleration control pattern Band the steering control pattern Sare selected, the vehicle control is executed in accordance with the deceleration control pattern Band the steering control pattern S.

10 10 When the server receives the control result information, the server learns a method for estimating the control occupant damage D1 based on the acceleration generated during the vehicle control and the posture change of the occupant during the vehicle control. Similarly, the server learns a method for estimating the collision occupant damage D2 based on the actual collision pattern and the posture change of the occupant at the time of collision. Furthermore, the server learns a method for estimating the object damage D3 based on the actual collision pattern and the actual damage to the collision object. The server transmits data relating to these learned methods to the present apparatus, and the present apparatusestimates the control occupant damage D1, the collision occupant damage D2, and the object damage D3 using the received data.

(First Modified Example) For example, when priority is given to the safety of the occupant, the control occupant damage D1 and the collision occupant damage D2are regarded as more important than the object damage D3. In such a case, as one example, Ga1 and Ga2 are set to “1.0,” and Ga3 is set to “0.5.”

(Second Modified Example) In the above embodiment, both the deceleration control pattern and the steering control pattern are prepared in advance; however, either one of the deceleration control pattern or the steering control pattern may be prepared in advance.

10 10 1 FIG. The present apparatuscan be applied to (or installed in/on) an engine vehicle, a hybrid electric vehicle (HEV), a plug-in hybrid electric vehicle (PHEV), a fuel cell electric vehicle (FCEV), and a battery electric vehicle (BEV). Furthermore, the present apparatuscan be applied not only to the vehicle (bus) shown in, but also to a passenger car and the like.